Longitudinal failure analysis using obtained sensor data of a physical machine

Abstract

An automated maintenance, monitoring and diagnostics infrastructure can include small, portable sensor devices, which can be attached to industrial machines and physical equipment. The sensor devices include wireless communication facilities and an accelerometer, capable of measuring vibration signals. Based on received sensor data, the system determines a condition state of the physical equipment and periodically evaluates additional data generated by the sensor device and determines an occurrence in a change of the condition state.

What is claimed is: 
     
         1 . A computer-implemented method comprising:
 receiving from a sensor device that is attached to a physical equipment, data generated by the attached sensor device, wherein the attached sensor device is configured to periodically generate data and transmit the generated data to a computing device, and wherein the generated data includes a timestamp for each respective occurrences of the generated data;   periodically receiving, via one or more servers, the generated data by the attached sensor device;   storing the generated data in a data repository the generated data, wherein the data repository includes generated data obtained from multiple different sensors each attached to a respective different physical equipment;   determining based on the periodically received generated data, the occurrence of one or more anomalies associate with the physical equipment;   determining a condition state associated with the physical equipment, wherein the condition state is determined by one or more trained machine learning models that receive as an input the generated periodically received generated data;   generating a first user interface depicting the determined one or more anomalies and the condition state, and depicting detailed information associated with an input selection, via the first user interface, of one of the one or more anomalies;   based on a change of the determined condition state, generating a configuration file and/or a command for transmission to the attached sensor device;   receiving by the attached sensor device, the generated configuration file and/or command;   based on the received generated configuration file and/or command, changing a sampling parameter of the attached sensor device to control the operation of the attached sensor device operation, wherein the sampling parameter includes any one of a sampling interval, a sampling frequency, a sampling rate, a sampling range or a sampling resolution;   operating the attached sensor device according to the changed sampling parameter;   periodically receiving, by the one or more servers, additional data generated by the attached sensor device;   evaluating, at least in part, the additional data generated by the attached sensor device, to determine that the condition state associated with the physical equipment has changed to a different condition state, wherein the evaluation is performed by the one or more trained machine learning model that receive as an input the additional data generated by the attached sensor device; and   generating a second user interface depicting the determined one or more anomalies, and depicting information indicative of the different condition state.   
     
     
         2 . The computer-implemented method of  claim 1 , wherein the change of the condition state is a change from an initial observe state to a worsening state indicating a gradual worsening of the determined one or more anomalies. 
     
     
         3 . The computer-implemented method of  claim 2 , wherein the change of the condition state is a change from the worsening state to the observe state indicating a stabilization of the determined one or more anomalies. 
     
     
         4 . The computer-implemented method of  claim 1 , wherein the change of the condition state is a change from a worsening state to a danger state indicating a continued worsening of the determined one or more anomalies. 
     
     
         5 . The computer-implemented method of  claim 4 , wherein the change of the condition state is a change from the danger state to the observe state indicating a significant improvement of the determined one or more anomalies. 
     
     
         6 . The computer-implemented method of  claim 1 , wherein the change of the condition state requires that a predetermined amount of time has passed since the condition state was determined in order to change from the condition state to the different condition state. 
     
     
         7 . The computer-implemented method of  claim 6 , wherein the change of the condition state requires that a predetermined threshold confidence level has been established that the condition state is actually changing to the different condition state. 
     
     
         8 . The computer-implemented method of  claim 4 , wherein the change of the condition state is a change from the danger state to the observe state indicating a significant improvement of the determined one or more anomalies. 
     
     
         9 . The computer-implemented method of  claim 1 , wherein over a period of time, the condition state changes from an initial condition state to a second condition state, the second condition state changes to a third condition state, and the third condition state changes back to the second condition state, and the second condition state changes to the initial condition state. 
     
     
         10 . The computer-implemented method of  claim 1 , wherein the attached sensor device is configured to operate in a hibernation mode to reduce battery consumption and conserve battery power, and wherein the attached sensor device is configured to periodically exit the hibernation mode according to the received generated configuration file and/or command. 
     
     
         11 . A physical equipment monitoring system, comprising:
 multiple sensor devices, each attached to separate physical equipment, wherein each of the multiple sensor devices comprise an accelerometer and a thermal sensor, and wherein each attached sensor device is configured to periodically generate data and transmit the generated data to a computing device, and wherein the generated data includes a timestamp for each respective occurrences of the generated data, and wherein the sensor device is configured to receive remote configuration files and/or command to change operational functionality of the sensor device; and   one or more servers, comprising one or more non-transitory computer-readable media storing computer-executable instructions that, when executed on one or more processors, cause the one or more processors to perform operations:   receiving, via the one or more servers, from each of the multiple sensor devices, data generated by the attached sensor devices;   storing the generated data in a data repository, wherein the data repository includes generated data obtained from the multiple different sensors;   determining, via the one or more servers, based on periodically received generated data for a first sensor device of the of the multiple sensor devices, the occurrence of one or more anomalies associate with a first physical equipment being monitored by the first sensor device;   determining, via the one or more servers, a condition state associated with the first physical equipment, wherein the condition state is determined by one or more trained machine learning models that receive as an input, at least a portion of the periodically received generated data generated by the first sensor device;   generating, via the one or more servers, a first user interface depicting the determined one or more anomalies and the condition state, and depicting detailed information associated with an input selection, via the first user interface, of one of the one or more anomalies; and   based on a change of the determined condition state, generating a configuration file and/or a command for transmission to the first sensor device;   wherein the first sensor device is configured to perform the operations of:
 receiving by the first sensor device, the generated configuration file and/or command; 
 based on the received generated configuration file and/or command, changing a sampling parameter of the first sensor device to control the operation of the first sensor device operation, wherein the sampling parameter includes any one of a sampling interval, a sampling frequency, a sampling rate, a sampling range or a sampling resolution; and 
 operating the first sensor device according to the changed sampling parameter; 
   wherein the one or more processors further perform the operations of:
 periodically receiving, by the one or more servers, additional data generated by the first sensor device; 
 evaluating, at least in part, the additional data generated by the first sensor device, to determine that the condition state associated with the first physical equipment has changed to a different condition state, wherein the evaluation is performed by the one or more trained machine learning model that receive as an input the additional data generated by the first sensor device; and 
 generating a second user interface depicting the determined one or more anomalies, and depicting information indicative of the different condition state. 
   
     
     
         12 . The monitoring system of  claim 11 , wherein the change of the condition state is a change from an initial observation state to a worsening state indicating a gradual worsening of the determined one or more anomalies. 
     
     
         13 . The monitoring system of  claim 12 , wherein the change of the condition state is a change from the worsening state to the observe state indicating a stabilization of the determined one or more anomalies. 
     
     
         14 . The monitoring system of  claim 11 , wherein the change of the condition state is a change from a worsening state to a danger state indicating a continued worsening of the determined one or more anomalies. 
     
     
         15 . The monitoring system of  claim 14 , wherein the change of the condition state is a change from the danger state to the observe state indicating a significant improvement of the determined one or more anomalies. 
     
     
         16 . The monitoring  system of 11 , wherein the change of the condition state requires that a predetermined amount of time has passed since the condition state was determined in order to change from the condition state to the different condition state. 
     
     
         17 . The monitoring system of  claim 16 , wherein the change of the condition state requires that a predetermined threshold confidence level has been established that the condition state is actually changing to the different condition state. 
     
     
         18 . The monitoring system of  claim 14 , wherein the change of the condition state is a change from the danger state to the observe state indicating a significant improvement of the determined one or more anomalies. 
     
     
         19 . The monitoring system of  claim 11 , wherein over a period of time, the condition state changes from an initial condition state to a second condition state, the second condition state changes to a third condition state, and the third condition state changes back to the second condition state, and the second condition state changes to the initial condition state. 
     
     
         20 . The monitoring system of  claim 11 , wherein each of the first sensor is configured to operate in a hibernation mode to reduce battery consumption and conserve battery power, and wherein the first sensor is configured to periodically exit the hibernation mode according to the received configuration file.